Western blot/Dot blot
Dot blot: Fibronectin (Fn) (1mg/ml), purified from human plasma by using gelatin-sepharose chromatography, was incubated with 10 mM oxidized glutathione (GSSG) (Sigma Aldrich) (volume ratio of Fn and GSSG is 4:1) at different temperatures: 37°C, 55°C, 75°C for 1 hour, and at 95°C for 20 minutes. Non-reducing Laemmli buffer with 20 mM maleimide (Sigma Aldrich) was then added and incubated at 95°C for 10 minutes. For the reversibility assay (de-glutathionylation), 50 mM 1,4-Dithiothreitol DTT (Sigma Aldrich) was incubated with all the samples at 95°C for 15 minutes. Anti-glutathione antibody (1:1000) (mAb100, Virogen) was used to detect glutathionylation.
Western blot: 10mM GSSG was added to 1mg/ml fibronectin sample and heated at 95°C for 20 minutes. Then non-reducing Laemmli buffer with 20 mM maleimide was mixed in at 95°C for 15 minutes. Anti-glutathione antibody (1:1000) (Virogen) and anti-fibronectin antibody (1:1000) (ab2413, Abcam) were used to detect glutathionylation and fibronectin.
Cell culture
Human foreskin fibroblasts (HFFs) were purchased from ATCC, passage 3 to 9 were used in all experiments. HFF cells were cultured in DMEM medium with 15% Fetal Bovine Serum (FBS) for production of cell-derived matrices, while other experiments (integrin switch and MRTF) were performed in serum free medium.
Single fiber deposition and decellularized matrix formation
All PDMS sheets were pretreated with 0.1M NaOH, 3% 3-aminopropyltriethoxysilane (APTES) and 1% glutaraldehyde before placing fiber and matrices.
Single fiber: Single Fn fibers were formed by previously described method32,33. Briefly, a single Fn fiber was formed by drawing out a single fiber from a drop of Fn solution. It was deposited on striated PDMS sheets (0.005” NRV G/G 40D 12”×12”) (Specialty Manufacturing Inc) fabricated by soft lithography with patterned ridges of 10×100μm and 50μm spacing. Masks (secondary modes) were made by photolithography.
Decellularized matrix: Human Foreskin Fibroblasts (HFF) cells were cultured on flat PDMS sheets (0.005” NRV G/G 40D 12”×12”) (Specialty Manufacturing Inc) with a seeding density of 10,000/cm2. After 7 to 9 days, HFF cells were lysed and removed with Latrunculin B (2µM) (Sigma Aldrich), EDTA (50mM) (Thermo Fisher Scientific), 2% sodium deoxycholate (DOC) and DNase I (1U/ml) (Amresco). Afterwards, the cell derived matrices are stretched to different strain levels (0% and 200% strain). Further details are described in31.
Protein (de-)glutathionylation and verification
Single Fn fibers were deposited on PDMS striated sheet from 1.5mg/ml Fn solution containing 5% Alexa 555-Fn (1.5mg/ml) and 95% non-labeled Fn (1.5mg/ml). For matrices, 48 hours before decellularization, 1µg/ml Alexa 555 labeled Fn was added into DMEM medium with 15% FBS which was then assembled by cell into matrix. Then single Fn fibers and decellularized matrices were subjected to the desired strain (0% and 200% strain) and labeled with 1mM FITC-GSSG (gift from Prof. Lisa Landino, The College of William and Mary) for an hour-long incubation at 37°C. For de-glutathionylation, 1mM DTT was added to the glutathionylated fibers or decellularized matrix, incubated for an hour at 37°C. Fluorescence images were taken with a UltraView Vox Spinning Disk Confocal Microscope (40X) (PerkinElmer) using Volocity 6.3.1 software for acquisition. Fluorescence signal ratio of FITC-SSG to fibronectin was used to determine (de)-glutathionylation signal strength within the 0% and 200% samples. Images of each condition were taken from different spot of same sample, each experiment was replicated three times.
AFM manipulation
All force data was collected by MFP-3D Bio (Asylum Research) paired with a TiEclipse inverted optical microscope (Nikon).
Single fiber mechanics: Single Fn fibers were deposited onto the striated PDMS substrate (as described before) mounted on a 50mm glass-bottomed Petri dish (FD5040-100, World Precision Instruments). Fn fibers were stretched (200%) with the underlying PDMS substrate, incubated with GSSG for glutathionylation, then relaxed to their original length. Each Fn fiber was pulled by an AFM tip (Model AC240TS-R3, Oxford instruments) with a spring constant of 2 N/m. The fiber was stretched at a rate of 320 nm/s. The strain at which the fiber breaks while pulling was reported as “extensibility”, and it is defined as , where . Young’s modulus data was determined by using the incremental stress-strain curve, Young’s modulus , where ɛ is the resulting strain and the applied stress , assuming fiber cross section is circular, . Fiber diameter is from AFM image with tapping mode. Details are described previously53. Fibers of each condition are from three independent samples. For single fiber series pulling: The AFM tip pulled a single Fn fiber incrementally and the force was collected. After the complete relaxation, fiber was then extended (kept at 200% strain) with the AFM tip to allow GSSG (10 mM for 30 minutes) or DTT (1 mM for 40 minutes) incubation followed by relaxation and subsequent force measurement as before. Replicated three times.
Decellularized matrix mechanics: A PDMS sheet with decellularized matrix (described in decellularized matrix above) was mounted on a 50mm glass-bottomed petri dish. Young’s modulus of matrix was measured with an AFM tip (Model MLCT-O10, Bruker) with a 25µm polystyrene bead (Polysciences, Inc.) glued onto tip C with a spring constant of 0.01 N/m. Cantilever was calibrated by thermal method. Force indentation curve was analyzed by the Hertz model, , where E is the elastic modulus, R is the radius of the bead on the tip, δ is the indentation, and ν is the Poisson’s ratio (in this experiment, 0.5). Measurements were made by indentation of 300 nm at a rate of 500 nm/s. Matrices of each condition are from three independent samples.
Immunofluorescence imaging
Imaging of Fn using H5 scFv: 48 hours before decellularization, 1 µg/ml Alexa 555 labeled Fn was added into DMEM medium with 15% FBS which was then assembled by cells into matrix. After decellularization, matrix was fixed with 4% paraformaldehyde and incubated in 3% Bovine Serum Albumin (BSA) in phosphate buffered saline (PBS) at room temperature for 1 hour. Then H5 (1:100) (10 µg/mL) antibody (Lab made antibody)31 was added to the surface of the decellularized matrix at 4°C overnight. Detection of H5 was enabled with anti-myc antibody (1:1000) (Fisher Scientific) and Alexa 647 secondary antibody (1:2000) (Thermo Fisher). Image analysis is described in31, briefly, ratiometric images were processed by custom Matlab code, Otsu’s method was used to threshold images, signal passed threshold on both Fn and H5 channel was masked, then the ratio of fluorescence intensity of H5 and Fn were compared among different groups. Images of each condition were taken from different spot of same sample, each experiment was replicated three times.
Integrin switch and MRTF: HFF cells were seeded at a density of 3,000/cm2 and cultured for 2 hours on decellularized matrix in serum free DMEM medium. Cells were then fixed by 4% paraformaldehyde for 10 minutes, permeabilized by 0.2% triton-X for 5 minutes. Samples were blocked with 5% normal goat serum before primary antibody incubation (1:200) overnight at 4°C.
Primary antibody used in these experiments: active αvβ3 (WOW-1, gift of Sanford Shattil, University of California, San Diego), active β1 (9EG7, BD Pharmingen), α-SMA PE conjugated (1A4, R&D Systems), anti-MKL1 (HPA030782, Sigma-Aldrich), rabbit anti-paxillin (Y113, Abcam).
All fluorescence images were imaged by a UltraView Vox Spinning Disk Confocal Microscope (PerkinElmer) (63X). Nuclei were stained with Hoechst (1:1000) (H3569, Thermo Fisher), while Alexa 488 phalloidin (1:40) (Thermo Fisher) was used for actin staining. For MRTF nuclear translocation experiment, z-stacks of 10 slices through the whole cell thickness were recorded. All image analysis were conducted by Volocity software. In Volocity, threshold was set accordingly to different channels for all the samples, only signal above the threshold were accounted for fluorescence intensity, then fluorescence intensity ratio of two channels (vary from different experiments) was used to compare. Cells of each condition are from same sample, each experiment was replicated three times.
Human Tissue Staining
Four lung cancer cryocores were provided by the Mid-Atlantic CHTN. We received a set of lung adenocarcinoma with patient-matched healthy control, and a set of bronchoalveolar carcinoma with patient-matched healthy control as well. Sections were cut on a cryotome at 10 um thickness. After thawing, the sections were fixed in 1% PFA for 10 minutes at room temperature (RT). After one wash in 1x PBS, they were permeabilized with 0.2% TritonX-100 in PBS for 10 minutes at RT. We blocked for 1 hour at RT with 5% BSA, 5% goat serum, 5% donkey serum and mouse Fc-block according to the provided protocol. Then the following antibodies were incubated overnight at 4C in PBST with 1% BSA: rabbit pAb anti-Fn (Abcam,1:1000 dilution, 647 channel); mouse anti-Glutathionylation (Virogen, 1:100 dilution, 555 channel).
The next day, after three washes in PBST of 10 minutes each, we incubated secondary antibodies for 1 hour at RT in PBST: goat anti-rabbit AF647 (Thermofisher), biotinylated donkey anti-mouse (abcam). After another set of washes, we incubated Streptavidin AF555 for 1 hour at RT. We then washed and added Prolong Gold+DAPI and allowed it to cure overnight. We intentionally left channel 488 open because tissue autofluorescence would provide structural information. DAPI in 405 channel stained cell nuclei. Slides were sealed and imaged on the same confocal microscope with the following settings:
Ch. 648: 48% laser power 400 ms exposure
Ch. 561: 37% laser power 300 ms exposure
Ch. 488: 37% laser power 200 ms exposure
Ch, 405 13% laser power 250 ms exposure
20 to 30 random 20X fields on each section were collected.
Unadulterated .tiff images were analyzed with our previously published31 ratiometric Matlab routine. The median pixel-by-pixel intensity ratio outputs were then used in Prism7 to produce the plots shown.
Mass Spec Sample Preparation
PDMS membranes with embedded, unmodified dECM and glutathionylated dECM (described above) were cut into small pieces (around 0.3 cm × 1 cm) and sonicated in 8M urea, 20mM HEPES for 30 minutes. For the unmodified dECM, the sample was centrifuged and the supernatant collected into a low-retention microfuge tube (Axygen) and reduced with Dithiothreitol (DTT; final concentration is 0.1M) by incubation at 80°C for 10 minutes. The sample was placed into a Vivacon centrifugal filter tube with 30 kDa cutoff (Sartorius) and centrifuged at 10K rpm for 15 minutes followed by alkylation with 300 μl 0.05 M iodoacetamide for 45 minutes in the dark. For glutathionylated dECM, reduction and alkylation were avoided to prevent loss of the modification. Both of the unmodified and glutathionylation dECM were washed three times with 300 μl 100mM ammonium bicarbonate at 10K rpm for 20 minutes each. Sequencing-grade chymotrypsin (Promega) was added to filter tube containing glutathionylated dECM, while trypsin (Promega) was added to unmodified dECM with the same enzyme to protein ratio (1:50). Both trypsin and chymotrypsin samples were incubated overnight while shaking at 350 rpm at 37°C and 25°C, respectively. Digested peptides were collected into fresh 1.5 ml low-retention microfuge tubes by centrifugation at 10K rpm speed for 5 min at room temperature and then desalted using a peptide desalting spin column (Pierce). Samples were then frozen solid at -80°C for at least 20 hours and sublimated to dryness via centri-vap (Labconco). Dried peptides were reconstituted by sonication in 50 μl 5% acetonitrile 0.1% formic acid for LC-MS analysis.
Mass spectrometry
Proteolytic peptides were analyzed by liquid chromatography mass spectrometry (LC-MS) using a Q-Exactive Plus Orbitrap mass spectrometer (Thermo Scientific) coupled with an UltiMate™ 3000 RSLCnano UPLC system (Dionex) equipped with an Acclaim Pep Map C18 RP column run in data-dependent acquisition mode with HCD fragmentation. Samples were separated by gradient LC with mobile phase A (2% ACN with 0.1% FA) and mobile phase B (80% ACN with 0.1% FA) with gradient from 4% B up to 90% B for over 150 min at a flow rate of 0.3 μl/min. The resulting MS RAW files were analyzed using the SEQUEST search algorithm via Proteome Discoverer 2.1 where glutathionylation (G) and oxidation (M) were included as variable modifications. MS fragmentation data were searched against the non-redundant UniProt database (https://www.uniprot.org/proteomes/UP000005640) using a peptide tolerance of 10 ppm in the MS1 scan and a fragment ion tolerance of 0.02 Da in the MS2 scan.
Mutant Fibronectin
Seven million HEK293 cells (ATCC) plated in a 10 mm dish in DMEM + 10% FBS were transfected using upscaled Lipofectamine 3000 reagents (Invitrogen). Specifically, 60 μl of lipofectamine 3000 reagent were combined with 40 μl of P3000 and 20 μg of mutant-Fn plasmid (kindly donated by Christopher Lemmon, Virginia Commonwealth University)10. The reagents were diluted in OptiMEM and administered to the cells according to the manufacturer protocol. Medium was replaced after 6 hours with DMEM +10% of Fn depleted FBS. Medium was collected, changed, and refrigerated every 2 days. Mutant-Fn was collected via gelatin-sepharose gravity filtration. After concentrating the eluted mutant-Fn, we obtained 250 μl at a concentration of 3.2 mg/ml.
Statistical analysis
GraphPad Prism was used to do all the analysis, data were presented as mean ± SD. Two-tail Mann Whitney test and Kruskal-Wallis test were used. Significant differences were considered when P < 0.05, *P < 0.05, **P < 0.01, ***P < 0.001, **** P < 0.0001.